2.1.1. CYP1A subfamily

The human CYP1A subfamily consists of two members, CYP1A1 and CYP1A2. CYP1A2 is mainly expressed in the liver, while CYP1A1 is primarily expressed in extrahepatic tissues. CYP1A1 is induced by cigarette smoking and polycyclic aromatic hydrocarbon (PAH) -type inducers in extrahepatic tissues, mainly the lung and placenta. In the liver, CYP1A1 is expressed at a very low level (Guengerich 1991; McKinnon et al. 1991; Pasanen & Pelkonen 1994; Raunio et al. 1995; Pelkonen et al. 1998). CYP1A1 has activity towards some PAHs (Guengerich 1991). The participation of CYP1A1 in drug metabolism in vivo is not considered here.

The constitutively expressed CYP1A2 in human liver is inducible by, for example, environmental compounds, such as constituents of cooked meat, cruciferous vegetables, cigarette smoke, PAHs and polychlorinated biphenyls (Adams et al. 1997). The metabolic activation of some arylamines, such as 4-aminobiphenyl, 2-aminoanthracene and 2-naphthylamine, is done by CYP1A2 (Guengerich 1991).

CYP1A2 represents about 15% of total CYP enzymes in the human liver (Pelkonen and Breimer 1994; Shimada et al. 1994). There is some variability between individuals in the CYP1A2 enzyme levels in the human liver (Shimada et al. 1994). Because of the large interindividual variation in CYP1A2 activities, the in vivo testing of this enzyme has been quite difficult to evaluate (Kunze and Trager 1993). CYP1A2 substrates among the currently used drugs include theophylline, caffeine, olanzapine, ondansentron, paracetamol, phenacetin and propranolol. Caffeine and theophylline have been considered in vivo diagnostic probes. The usually employed in vitro substrate is ethoxyresorufin (Table 2), which seems to be quite specific to CYP1A2 (Burke et al. 1985). The substrates and their enzyme kinetic parameters have recently been reviewed by Pelkonen et al. (1998).

Furafylline (Table 2) and fluvoxamine have been the most widely used in vitro diagnostic inhibitors. Furafylline is a potent and selective mechanism-based inactivator of CYP1A2 (Sesardic et al. 1990), but it has no function as an in vivo inhibitor because of its interactions with caffeine (Tarrus et al. 1987).

Fluvoxamine is also used as an in vivo diagnostic inhibitor for CYP1A2 (Brosen et al. 1993; Rasmussen et al. 1995). Fluvoxamine is known to inhibit moderately CYP2D6 and CYP2C19, though at a higher concentration range than CYP1A2 (Jeppesen et al. 1996), but not CYP1A1, CYP2A6, CYP2E1 or CYP3A4 in vitro. Venkatakrishnan and co-workers have shown that fluvoxamine inhibits CYP2C19 equally potently as CYP1A2 in lymphoblast-expressed human CYPs and, to a slightly lesser extent, CYP2C9 and CYP3A4 (von Moltke et al. 1995; 1996; Venkatakrishnan et al. 1999).

2.1.2. CYP2A subfamily

In humans, there are three genes in the CYP2A subfamily: CYP2A6 codes the enzyme catalysing coumarin 7-hydroxylation (about 10% of total P450), while the product of CYP2A7 is inactive and CYP2A13 is not expressed in the liver (Pelkonen et al. 2000). In the olfactory mucosa, CYP2A13 is highly expressed as a catalytically active protein (Gu et al. 2000). Recently, CYP2A6 has been reported to be polymorphically expressed in the human liver (Oscarson et al. 1998). No endogenous substrate for CYP2A6 has been found.

It has been recently shown that CYP2A6 participates in the metabolism of nicotine (Cashman et al. 1992; Nakajima et al. 1996b) and its metabolite cotinine (Nakajima et al. 1996a). This enzyme also catalyses the metabolic activation of several procarcinogens and promutagens, such as aflatoxin B₁ (Yun et al. 1991; Salonpää et al. 1993) and several nitrosamines (Yamazaki et al. 1992; Tiano et al. 1994). Some drugs and chemicals, including halothane and coumarin (Table 2) – which is widely used as a probe substance for CYP2A6 both in vitro and in vivo - are also metabolised by this enzyme (Rendic & Di Carlo 1997; Pelkonen et al. 1998).

Draper and coworkers (1997) have extensively studied the commonly employed P450 substrates and inhibitors with regard to their effect on CYP2A6-catalysed coumarin 7-hydroxylation. They also screened the effects of some commonly used solvents on the same reaction. They found that many of the studied chemicals were inhibitors of CYP2A6, although earlier presumed to be selective for some other CYPs. For example, ketoconazole (Table 2) is used as a CYP3A4-specific inhibitor, but it was also found to inhibit CYP2A6 quite potently. Also, many commonly used organic solvents inhibit coumarin 7-hydroxylation at substrate concentrations near K_m.

More substrates and inhibitors currently known to be metabolised by or to interact with CYP2A6 in vitro and in vivo have been summarised by Pelkonen et al. (2000).

2.1.3. CYP2B subfamily

CYP2B6 is the only active member of the CYP2B subfamily in man, although the CYP2B7 gene has also been found in the genome (Czerwinski et al. 1994). For a long time, it was thought that CYP2B6 would not be expressed in every human liver. For example, Mimura et al. (1993) assessed 50 human liver samples by immunoblotting analysis with antimonkey P4502B antibody for CYP2B6 and found only about of the livers to contain the protein (see also Yamano et al. 1989 and Shimada et al. 1994). The highest level of CYP2B6 found by Mimura and co-workers was <2% of the total P450. Today, several groups have shown that this enzyme is expressed in every human liver, but the levels of expression differ between individuals (Ekins et al. 1997; Stresser & Kupfer 1999). The amount of CYP2B6 protein seems to be below 1% of the total P450 present in the human liver, and the reason for the earlier reports about non-expressing individuals is probably the weak sensitivity of either the antibodies used or the detection methods. Today, there are very sensitive and specific anti-peptide antibodies available for quantitative immunological studies (Stresser & Kupfer 1999). Similarly to the other studies on CYP2B6, they showed the expression of this enzyme to be highly variable.

CYP2B6 has been postulated to be co-regulated with CYP3A4 or at least to be inducible by the same inducers, such as rifampicin and dexamethasone, as well as by phenobarbital in cultured human hepatocytes (Strom et al. 1996).

The catalytic activity of CYP2B6 has recently been under efficient research by many groups. Especially Ekins and coworkers (1997, 1998) have evaluated several chemicals and drugs as specific substrates for this enzyme. Among the suggested CYP2B6-catalysed reactions are O-deethylation of 7-ethoxy-4-trifluoromethylcoumarin (7-EFC), S-mephenytoin N-demethylation, stereoselective hydroxylation of RP 73401 [3-cyclopentyloxy-N-(3,5-dichloro-4-pyridyl)-4-methoxybenzamide]; to RPR 113406 (Stevens et al. 1997) and bupropion hydroxylation (Faucette et al. 2000). The availability of the other selective substrate presented above, RP 73401 and its CYP2B6-formed metabolite RPR 113406 prepared by Rhõne-Poulenc Rorer Co., is unknown. Recently, Kobayashi et al. (1999) evaluated the role of CYP2B6 in S-mephobarbital N-demethylation by using 10 cDNA-expressed human CYPs. They found that only CYP2B6 catalysed the reaction. Also, the activity correlated strongly with the immunodetectable CYP2B6 levels in microsomes from 10 individual human livers. This activity is probably quite usable if the substrate, S-mephobarbital, could be easily separated from its R-enantiomer. The metabolite, phenobarbital, is commonly used as a P450-inducing agent (Pelkonen et al. 1998).

Orphenadrine has been widely used as an inhibitor of CYP2B6, though its selectivity towards this isoform is questionable (Murray & Reidy 1990; Chang et al. 1993; Ekins et al. 1997; Guo et al. 1997). Although actively searched, no selective inhibitor for CYP2B6 has been found yet.

2.1.4. CYP2C subfamily

The CYP2C subfamily is the second most abundant CYP protein in the human liver, representing about 20% of the total P450 (Shimada et al. 1994). This subfamily consists of three active members in the human liver, namely CYP2C8, CYP2C9, and CYP2C19. Of these, CYP2C9 and CYP2C19 have been characterised to be polymorphically expressed (Goldstein & de Morais 1994; Gill et al. 1999).

CYP2C8 has been thought not to play an important role in drug metabolism since it is expressed at very low levels in the human liver. Still, the new, very potent anti-cancer drug, taxol, for example, is partly metabolised by CYP2C8 (Harris et al. 1994; Sonnichsen et al. 1995). CYP2C8 also participates in the metabolism of the endogenic agents retinol and retinoic acid (Leo et al. 1989). The involvement of this member of the CYP2C subfamily in the metabolism of NCEs has been difficult to study because there is no good, specific substrate and inhibitor for this isoform. In many cases, CYP2C8 also participates, to a small extent, in the metabolism of the CYP2C9 substrates (Wrighton et al. 1993b).

CYP2C9 is the major CYP2C isoform in the human liver (Goldstein & de Morais 1994), and it has been shown to be genetically polymorphic with at least three different alleles that produce differently active protein. The functional consequences of these polymorphisms are not yet clear, although CYP2C9 has a major role in the metabolism of many clinically important, weakly acidic drugs, such as S-warfarin (Rettie et al. 1992), tolbutamide (Table 2), phenytoin (Doecke et al. 1991), sulphamethoxazole (Cribb et al. 1995) and many of the non-steroidal anti-inflammatory compounds (Leeman et al. 1993) as well as the new drug in this class, a cyclooxygenase-II selective inhibitor, celecoxib (Tang et al. 2000). The frequencies of the two variant alleles, CYP2C9*2 and CYP2C9*3, have been reported to range from 7 to 19% in Caucasian populations (Furuya et al. 1995; Sullivan-Klose et al. 1996; Stubbins et al. 1996).

With CYP2C19, the genetic polymorphism leads to the poor metaboliser (PM) phenotypes exhibiting less active or completely inactive S-mephenytoin 4’-hydroxylase (Table 2). This PM phenotype is produced by at least two major and several minor variant alleles of CYP2C19 (Goldstein & de Morais 1994) and, consequently, CYP2C19 substrates are not metabolised as expected (Pelkonen et al. 1998). This may lead to accumulation of the drug and to in vivo concentrations exceeding the therapeutic level and producing unexpected toxic effects. The search for CYP2C19-selective inhibitor both in vitro and in vivo is undeer way, since omeprazole (Table 2), which is usually employed, also inhibits other CYPs (Funck-Brentano et al. 1997). The deficiency of the 4’-hydroxylation pathway of S-mephenytoin occurs in 2 to 5% of the Caucasian population (Relling et al. 1990).

2.1.5. CYP2D subfamily

The human genome includes only one functional gene in the CYP2D subfamily, namely CYP2D6 (Nelson et al. 1996). There also exist two CYP2D7 pseudogenes and two pseudogenes of CYP2D8 (Heim & Meyer 1992; Nelson et al. 1996). More than 60 allelic variants have been reported for CYP2D6 (Streetman et al. 2000; see: http://www.imm.ki.se/CYPalleles). Of these, most affect the activity of the expressed protein. CYP2D6 represents 1 to 5% of the total P450 (Pelkonen & Breimer 1994; Shimada et al. 1994; Pelkonen et al. 1998), and 7-8 % of the Caucasian population are PMs for this enzyme (Heim & Meyer 1992).

The search for endogenous substrates for CYP2D6 has recently become more interesting because it has been shown that CYP2D6 is associated with Parkinson’s disease (Barbeau et al. 1985; Armstrong et al. 1992; Smith et al. 1992; Agúndez et al. 1995) and with susceptibility to develop liver and lung cancer (Idle 1981; Ayesh et al. 1984; Agúndez et al. 1994; Caporaso et al. 1995; Bouchardy et al. 1996). Martínez et al. (1997) suggested a neurotransmitter, tryptamine, as an endogenous substrate for CYP2D6. The metabolism of tryptamine into tryptophol was catalysed in a NADPH-dependent manner and inhibited by a specific CYP2D6 inhibitor, quinidine (Table 2).

One common feature of CYP2D6 substrates is that they contain at least one basic nitrogen atom at a distance of 5 or 7Å from the oxidation site. Secondly, there is a planar hydrophobic area near the oxidation site, and thirdly, the substrates exhibit a negative molecular electrostatic potential above the planar part of the molecule (Koymans et al. 1992; Strobl et al. 1993; de Groot et al. 1997). The drugs known to be substrates for this enzyme include antiarrhythmic and other cardiovascular drugs, β -adrenergic blocking agents, tricyclic antidepressants, neuroleptics and many other commonly used therapeutical agents (Cholerton et al. 1992). The mostly used in vitro model reactions are dextromethorphan O-demethylation (Table 2), debrisoquine 4-hydroxylation and bufuralol 1’-hydroxylation. Of these substrates, debrisoquine is also employed in in vivo studies as a CYP2D6 model substance, although its use is becoming more difficult because it is not marketed any more (Pelkonen et al. 1998, and references therein).

As a reference inhibitor for CYP2D6, quinidine is widely used in drug metabolism studies. It is a specific and potent inhibitor with a K_i value of 0.06 µM for CYP2D6, and the CYP next in sensitivity, CYP3A4, has a K_i value around 10 µM (Broly et al. 1989; Bourrie et al. 1996).

Due to the fact that numerous drugs on the market are metabolised by CYP2D6 and the polymorphic nature of the expression of this protein, it is strongly suggested that the affinities of the NCEs on this CYP form should be characterised to predict and to possibly avoid drug-drug interactions.

2.1.6. CYP2E subfamily

The coding region of CYP2E1 is highly conserved in different ethnic groups and species. There are significant interethnic differences in polymorphisms, but there is no clear evidence of any of the reported polymorphisms to be related to altered in vivo activities. All the reported polymorphisms are found in the noncoding regions of the gene, suggesting an important role in the biotransformation of endogenous substances (Ronis et al. 1996; Yin et al. 1997).

The CYP2E1 expression is regulated by many factors. For example, fasting elevates the level of the protein by increasing the transcription of the gene (Johansson et al. 1990) and diabetes by stabilising the mRNA (Song et al. 1987), while isoniazid increases the translation efficiency (Park et al. 1993) and affects the enzyme stabilisation similarly to ethanol and imidazole (Eliasson et al. 1990).

The CYP2E1 substrates include very few clinically important drugs. Actually, only paracetamol (Patten et al. 1989), caffeine (Gu et al. 1992), chlorzoxazone (Peter et al. 1990; Table 2) and enflurane (Thummel et al. 1993) are worth mentioning here. Most organic solvents and anesthetics, short-chain alcohols and many nitrosamine and azo carcinogens (Hong & Yang 1985; Koop 1992; Sohn et al. 1991; Yang et al. 1985) are also xenobiotic substrates of this enzyme. CYP2E1 participates in the metabolism of many endogenous substances, such as lipid peroxidation products (Terelius & Ingelman-Sundberg 1986), ketones (Koop & Cassazza 1985) and fatty acids, such as linoleic and arachidonic acids (Laethem et al. 1993). A number of other small-molecular CYP2E1 substrates are listed in the review article by Ronis et al. (1996).

Pyridine (Table 2) is a specific in vitro inhibitor and inhibits CYP2E1 at relatively low concentrations (Hargreaves et al. 1994; Taavitsainen et al., unpublished results). The metabolite of disulfiram, diethyldithiocarbamate, has also been used as an in vitro inhibitor (Guengerich et al. 1991; Brady et al. 1991), though its selectivity is questionable (Yamazaki et al. 1992).

2.1.7. CYP3A subfamily

The CYP3A subfamily represents about 30% of the total P450 content in the human liver (Shimada et al 1994; Pelkonen and Breimer 1994), although the levels of the protein may vary 40-fold among individuals (Guengerich 1995a). This subfamily consists of three members (Nelson et al. 1996): CYP3A4 is the most abundant CYP enzyme in the human liver and it is expressed in several tissues, but the expression in the liver and in the small intestine is of major interest in view of the metabolism of drugs and other xenobiotic chemicals (Guengerich 1999), CYP3A5 is a minor, polymorphic CYP3A enzyme, which is expressed in the lungs (Kivistö et al. 1996; Anttila et al. 1997) and the colon (Gervot et al. 1996). About 20% of individual adults express CYP3A5 at a high level in the liver (Guengerich 1999). CYP3A7 is expressed in the fetal liver and in the adult endometrium and placenta (Schuetz et al. 1993; Hakkola et al. 1994).

The members of the CYP3A subfamily have overlapping catalytic specificities, but some selectivity exists. CYP3A4 participates in the metabolism of about half of the drugs in use today (Bertz and Granneman 1997). For example, testosterone 6β -hydroxylation (Table 2), midazolam 1’- and 4-hydroxylations (Table 2), nifedipine oxidation, and erythromycin N-demethylation are catalyzed by this enzyme. The known substrates of CYP3A4 vary in size from small molecules, such as acetaminophen (M_r 151), to cyclosporin A (M_r 1201) (Guengerich 1999). In addition to the substrates listed above, we could further mention physiologically important progesterone and andostenedione (Waxman et al. 1991), cortisol, quinidine, diltiazem, lidocaine, lovastatin, troleandomycin, warfarin, and triazolam (Guengerich and Shimada 1991; Wrighton and Stevens 1992).

CYP3A is inducible by many drugs, for example, rifampicin, dexamethasone, carbamazepine and phenobarbital type inducers (Pelkonen et al. 1998). The induction of CYP3A has an effect on interindividual variation and affects both bioavailability and drug-drug interactions (Guengerich 1999).

The inhibitors of CYP3A have a wide variety of chemical structures. For example, azole-type fungicides, ketoconazole and itraconazole are potent inhibitors. Ketoconazole (Table 2) also inhibits other CYPs than CYP3A4, but at a concentration of 1 µM it is relatively selective for CYP3A4 (Newton et al. 1994; Baldwin et al. 1995). Gestodene, a progesterone analog with a steroid structure has been long known as a mechanism-based CYP3A inhibitor (Pelkonen et al. 1998). Gestodene is also a substrate for CYP3A4.

The substrate specificity and catalytic features of CYP3A4 have recently been a target of active research. It has been shown by many groups that the properties of CYP3A4 have not yet been thoroughly elucidated (Ueng et al. 1997; Korzekwa et al. 1998; Wang et al. 2000). For example, the role of cytochrome b₅, divalent cations and the rate-limiting steps in the catalysis are still under study (Guengerich 1999). Wang et al. (2000) studied the in vitro drug-drug interaction patterns with four commonly used CYP3A4 substrates. The results indicated that CYP3A4 is a very complex enzyme and that its interaction patterns are substrate-dependent.

Thanks to the unique properties of CYP3A4, the enzymatic processes catalysed by CYP3A4 do not always follow the typical competitive inhibition kinetics. A substrate can either inhibit or stimulate the in vitro metabolism of another substrate, or activate its own metabolism (Shou et al. 1994). The kinetics can be either cooperative or allosteric, depending of the binding sites of the two substrate/inhibitor molecules or one molecule of two substrates each or one molecule of the substrate and an effector (Shou et al. 1994; Ueng et al. 1997; Korzekwa et al. 1998; Guengerich et al. 1999; Wang et al. 2000).

Although the cooperativity in CYP3A4-catalysed reactions has been known for a long time, there has not been any major progress in this field. The great majority of studies suggest that there are at least two distinct binding sites for the substrate and the effector. The location of the effector binding site – whether at the active site or at a separate allosteric site – is not yet known (Ueng et al. 1997). Harlow and Halpert (1998) presented a model for cooperativity which suggests that the substrate and effector molecules bind at adjacent sites and are both part of a large binding cavity. This model agrees with the one presented by Shou et al. (1994), with the exception that they suggested two catalytic sites both with access to the reactive oxygen.

Based on these results, it is concluded that the active site of CYP3A4 is very large and flexible, which postulation is supported by the wide substrate selectivity. Shou and co-workers (1994) also suggested that the activator, when binding to the active site, excludes water molecules from the substrate cavity and prevents hydrogen peroxide release. These models combined could together explain the atypical CYP3A4 kinetics, including activation, autoactivation, mutual inhibition, partial inhibition, substrate inhibition, biphasic saturation curves and alteration of regio-specificity.

Still, although much effort has been invested in CYP3A4 research, there may still be many properties of this enzyme that need to be studied. One should bear in mind the likelihood that because CYP3A4 plays such an important role in xenobiotic metabolism, its catalysis may not be very simple in nature. The conclusions drawn on the basis of only one substrate or inhibitor should be regarded as tentative.